CMR Tools

Segment has primarely been developed for cardiac MR (CMR) and hence there are a number of quantification tools available for cardiac MR. Click on tabs below to learn more about Segment capabilities for CMR.

LV Segmentation

The key to global and regional LV function is a good and reliable segmentation. The segmentation engine is a unique and fast truly three-dimensional time resolved algorithm based on the concept of deformable models. The algorithm uses image information from all slices and time frames simultaneously. The unique features of the Segment LV segmentation algorithm are:

  • Truly 3D+T segmentation
  • Very fast (takes less than 20s for 12 slice and 30 time frames on a fast PC)
  • Image modality independent (fixed parameter sets)
  • Ease of manual interaction by dynamic contour editing and placing pins
  • Unique method of incorporating long-axis motion
  • Unique possibility to choose wether to include or exclude the papilaries. This option is crucial to good regional wall motion analysis

From the left ventricle segmentation the following global LV parameters are directly available:

  • Myocardial mass
  • End systolic volume
  • End diastolic volume
  • Stroke volume
  • Cardiac output
  • Ejection fraction
  • Peak ejection rate
  • Peak filling rate
  • Volume curve
  • Volume loop
  • 3D visualization of the left ventricle

RV Segmentation

In Segment there are several semi-automated tools that can be used to Segment the RV endocardium. The tools are semi-automated, but require more user interaction compared to the left ventricle tools. The figure below shows an example of RV segmentation in end-diastole.

Viability Analysis – LGE

The viability or late gadolinium enhancement(LGE) detection in Segment is a unique algorithm that compensates for partial volume effects in a weighted approach. The algorithm is the only general available algorithm that is validated in an experimental setting. The algorithm is validated experimentally in a pig model, in computer phantoms and in 40 patients where the mean of three observers were used as gold standard. The bias or the variability of the algorithm did not differ from manual observers.

Strain Analysis

The strain analysis algorithm in Segment is carefully validated both in human and phantom data. The analysis is performed in velocity encoded longaxis MR images and quantifies both radial and longitudinal strain.

Flow Analysis

The flow analysis algorithms in Segment are carefully validated agains timer and beaker measurements, and also digital computer phantoms. The flow analysis option allows unlimited number of ROI’s. An automatic tracking of vessels is available for drawing ROI’s which minimizes user interaction. Each ROI can be labelled and colored individually.

You can plot the following parameters

  • Volume flow curves
  • Backward and forward flows
  • Mean velocity curves + standard deviation
  • Vessel area curves
  • Vessel diameter curves
  • Signed kinetic energy

All parameters are directly exportable to Excel for easy statistical analysis.

From the flow analysis interface it is also possible to view 3D profiels of the flow. This allows a rapid quantification and understanding of skewness of flow profiles.

There is also a tool for automated phase unwrapping if you should have set the VENC incorrect. This have the ability to save data sets that otherwise would be impossible to quantify. The phase unwrapping may either be performed automatically or by manual user interactions.

For flow analysis a module for automatic eddy current compensation is implemented. This module allows for linear and polynomial phase background subtractions. There is also a mode for compensating eddy currents by usage of stationary phantom that is placed in the scanner after the patient has left.